The etiology of rheumatoid arthritis (RA) is unknown, but considerable interest in a possible involvement of T cells cross-reactive with mycobacterial 65kDa heat shock protein (hsp-65) and self antigens of the joint has developed during recent years. It has never been satisfactorily explained why cross-reactive response to hsp-65 are likely to be injurious to self; shouldn't the individual be tolerant to the shared determinants? We would like to explain athe immune response to shared epitopes of hsp 65 in the context of 'cryptic' and 'dominant' determinants of hsp-65. Our hypothesis is that shared determinants on human hsp-65. Our hypothesis is that shared determinants on human hsp-65 in fact are responsible for induction of rheumatoid arthritis, but that there has to exist a particular constellation among the cross-reactive (CR) determinants on the bacterial vs. the mammalian hsp's. The CR determinants must be dominant on the bacterial hsp to initiate a host response, while they have to be cryptic on the self-hsp. The crypticity is necessary on the self protein so that one is not tolerant to these determinants, allowing potentially self-reactive T cells against the cryptic determinants to be stimulated when the microbial hsp arrives. Otherwise there would be no self-reactive response, but only one directed against microbial determinants. The clear prediction of this hypothesis is that pathogenic determinants would be cryptic ont he self antigen.